Compass becomes crazy (SOLVED)

[seandepagnier]

user is "tc" not "pi"

[Daniel Simon]

Hello,
I moved the tinypilot farther from the main floating compass, now it is 1.6m far. After circling and one afternoon sailing with few wind and swell, compass calibration unlocked,  I got a stable calibration close to the previous one, with a 202 remaining value for the Y bias...

Otherwise no other problems with the pilot, but the conditions were quiet!

[seandepagnier]

I marked as solved, because there is a workaround (locking)

I have made changes to pypilot so that it will not calibrate when the pilot enabled (it can still be locked and wont calibrate even in standby). This is a bit unfortunate but I think probably best until the calibration is improved to not converge incorrectly from rough conditions.

[Daniel Simon]

Thanks.

[Jean-Marc Douroux]

Your last tynipilot image allowed me to find exactly the same behavior with the same setting as before. The problem of the heading not being very precise due to the calibration was ultimately the cause of the difficulties in finding a correct setting. Everything works perfectly again with my old settings.
the problem was probably amplified because I use a significant gain DD. It had appeared since I had installed an image with several tuning sets.
Thanks Sean